Memory problem in multithreading code with Python

I am writing a code to obtain the size of all Azure DataLake Storage files. I'm using a multithread process to reduce the execution time of the code.

The code does the following:

1. It get the first 3 levels folders of my datalake:

• raw/project1/2020/01
• raw/project1/2020/02
• ...
• raw/projectN/2020/01

2. It creates 4 threads
3. It put each folder in a Queue
4. For each folder I obtain all the files and the sizes in the folder
5. It writes each folder result in a file
6. It upload the file to a StorageAccount

    def downloadFolder(adls_path):
        #return 
        try:
            return getAllSubdirectories(adls_client, adls_path)
        except PermissionError:
            #write folder into file

    class DownloadWorker(Thread):

        def __init__(self, queue):
            Thread.__init__(self)
            self.queue = queue

        def run(self):
            while True:
                directory = self.queue.get()
                try:
                    data = downloadFolder(directory)
                    if data is not None:         
                        upload_file_to_blob(sa_tenant, sa_client, sa_secret,
                                            sa_name, sa_container,
                                            file_path)
                    else:
                        print(Empty folder)
                finally:
                    self.queue.task_done()

    if __name__ == "__main__":
        adls_client = connectADLS(
            sa_tenant(), sa_client, sa_secret, adls_name
        rootFolders = getSubdirectories(adls_client, adls_root_dir)
        random.shuffle(rootFolders)
        queue = Queue()
        for i in range(4):
            worker = DownloadWorker(queue)
            worker.daemon = True
            worker.start()
        for Folder in rootFolders:
            queue.put((Folder))
        queue.join()
        sys.exit(0)

Monitoring the memory with memory_profiler I see an increase of the memory that never falls:

Why I am using 2000 MiB of memory if the biggest file created is over 300 MB and how can I control this memory usage?